PROJECT SUMMARY Treatment of advanced melanoma remains a major challenge despite recent FDA approvals of targeted therapies and immunotherapies that significantly improve patient survival. Immune checkpoint inhibitors, e.g., anti-CTLA-4 (ipilimumab) or anti-PD-1/PD-L1 (nivolumab), augment anti-tumor immunity to improve clinical outcomes in certain patients. Concurrent ipilimumab and nivolumab therapy is more efficacious than immune checkpoint monotherapy, albeit at the expense of increased toxicity. Hence, there is a need for novel treatment combinations to boost efficacy of immune checkpoint inhibitors without additional toxicity, and to identify robust biomarkers to predict which melanoma patients are most likely to benefit from immunotherapy. Increasing evidence supports a complex interplay between tumor angiogenesis and anti-tumor immunity, largely through vascular endothelial growth factor (VEGF) signaling. Thus, combining the inhibition of angiogenesis with immune checkpoint blockade provides a rational strategy to enhance anti-melanoma immunity. Indeed, a recent phase I study of ipilimumab and bevacizumab (anti-VEGF) in melanoma patients demonstrated better clinical activity than ipilimumab alone. Moreover, tumor biopsies from patients treated with the combination showed elevated CD8+ T cell infiltration and endothelial cell activation, suggesting that this combined treatment approach might improve melanoma immunotherapy efficacy. However, a critical obstacle to this strategy is the lack of a biomarker to predict which immunotherapy- treated patients would benefit from angiogenesis inhibition. Our preliminary and recently published data identified a KDR (VEGFR2) germline variant (Q472H) in ~35% of melanoma patients in two independent cohorts. We showed that: (i) KDR Q472H is associated with increased tumor microvessel density (MVD) and VEGF secretion; (ii) KDR Q472H melanoma cell growth and invasion are sensitive to VEGFR2 inhibition, supporting angiogenesis blockade as a rational therapeutic strategy in patients carrying this variant; (iii) increased ERK1/2 and Akt signaling in KDR Q472H melanoma cells compared with wild-type (WT) cells, and (iv) a pilot cohort of melanoma patients who did not respond to anti- CTLA-4 therapy had increased KDR Q472H prevalence. Our working hypothesis is that germline KDR Q472H is a functional determinant of melanoma angiogenesis (and thus anti-tumor immunity), and might dictate the clinical response to anti-CTLA-4 or anti-PD-1 therapy. We will utilize melanoma patient tissues with extensive clinicopathological annotation, germline DNA from three independent immunotherapy cohorts, and a transgenic melanoma mouse model established by our group to address these questions. The direct translational impact of our work is that germline KDR Q472H might (i) predict patient benefit from combined anti-angiogenesis and immune checkpoint blockade, and (ii) provide an actionable target to sensitize tumors to these therapies.